Re: [RFC] Merge criteria for tree-ssa

[Paul Brook <paul at codesourcery dot com>]

> In message <http://gcc.gnu.org/ml/gcc/2004-02/msg00506.html>
> Bud Davis <bdavis9659 at comcast dot net> wrote
>
> > The criteria that makes most sense from a users viewpoint is that
> > tree-ssa could not be merged until gfortran was as good as g77 at
> > compiling f77.

> In message <http://gcc.gnu.org/ml/gcc/2004-02/msg00507.html>,
> Mark Hahn <hahn at physics dot mcmaster dot ca> wrote
>
> > as another user, I strongly disagree.  can't people with f77 code
> > simply compile with older versions of gcc?
>
> Not if they need (for example) both
> (a) interlanguage calling between C, C++, and F77 code, which to be
>     safe means (among other things) using the same gcc versions for
>     all three languages, *and*
> (b) the excellent C++ support of newer gcc versions

The problem is that g77, and by implication (a), is incompatible with:

(c) Improved optimizations provided by the tree-ssa infrastructure. These seem 
to be particulary beneficial for the C++ code in (b).

This is the primary purpose of the tree-ssa branch. The (mostly temporary) 
Fortran breakage is an unfortunate side-effect. The inconvenience for Fortran 
77 users has to be balanced against the gains for other users.

I expect there are also people who need to mix C/C++ and F90 code. This is 
impossible with current mainline.

Paul